It has been a significant week for AI startup Anthropic, the creator of the Claude family of large language models (LLMs) and chatbots.
On Monday, January 22nd, Anthropic reported that a contractor unintentionally sent a file containing non-sensitive customer information to a third party. This file included a "subset" of customer names and open credit balances as of the end of 2023.
“Our investigation indicates this was an isolated incident due to human error — not a breach of Anthropic systems,” stated an Anthropic spokesperson. “We have informed affected customers and provided them with appropriate guidance.”
This incident coincided with an announcement from the Federal Trade Commission (FTC), which is examining Anthropic's strategic partnerships with Amazon and Google, in addition to OpenAI's affiliations with Microsoft. Anthropic clarified that the contractor's error is unrelated to the FTC investigation, on which they did not comment.
Accounts Information Misdirected
Windows Report obtained a screenshot of an email sent by Anthropic to customers regarding the information leak caused by a third-party contractor. The email disclosed the "account name and accounts receivable information as of December 31, 2023." Here is the email content:
Important alert about your account:
We want to inform you that one of our contractors inadvertently misdirected some accounts receivable information from Anthropic to a third party. The information included your account name and accounts receivable details as of December 31, 2023, indicating an open credit balance. Importantly, this data did not contain sensitive personal information, such as banking or payment details. Our investigation has determined that this incident was due to an isolated error, with no compromise of our systems. We have seen no indication of malicious intent related to this disclosure.
Anthropic urged customers to remain vigilant against any suspicious communications that appear to come from the company, such as unusual payment requests or unexpected emails.
“We sincerely regret this incident and any disruption it may have caused you. Our team is ready to support you,” added the company.
Only a Subset of Users Affected
In response to inquiries about the leak, Anthropic confirmed that only a "subset" of users was impacted but did not disclose specific numbers. As data breaches remain alarmingly frequent—95% attributed to human error—this incident underscores the anxiety organizations face when using third-party LLMs like Claude for proprietary data.
The leak has heightened concerns among enterprises that sensitive data could be compromised through third-party LLMs. Notably, Samsung banned ChatGPT last spring after sensitive company data was leaked.
Regulatory Scrutiny on AI Partnerships
Since its founding in 2021, Anthropic has experienced rapid growth, boasting a valuation of $18.4 billion and significant funding efforts, including up to $2 billion from Google and another $4 billion from Amazon. The company also aims to secure an additional $750 million round led by Menlo Ventures.
However, Anthropic’s partnerships with AWS and Google have attracted attention from the FTC, which recently issued 6(b) orders to Amazon, Microsoft, OpenAI, Anthropic, and Alphabet for detailed disclosures regarding their strategic collaborations.
Key partnerships under scrutiny include:
- Microsoft and OpenAI's extended collaboration announced on January 23, 2023
- Amazon and Anthropic's strategic partnership launched on September 25, 2023
- Google's expanded AI alliance with Anthropic, established on November 8, 2023
The FTC has requested analysis on how these associations could affect market competition and has noted that both the European Union and the United Kingdom are reviewing Microsoft’s AI investments as well.
FTC Chair Lina Khan stated, “We are investigating whether these relationships allow dominant firms to wield undue influence, potentially compromising fair competition."
Anthropic’s Strategic Collaborations
Anthropic has partnered with AWS and Google since its establishment, significantly expanding these relationships. Amazon plans to invest up to $4 billion, gaining minority ownership in Anthropic and serving as its primary cloud provider.
Anthropic has also committed to delivering future generations of its AI models to AWS customers via Amazon Bedrock, providing early access to model customization features.
Additionally, Anthropic utilizes Google Cloud services, integrating tools such as Google’s TPU v5e for its Claude LLM and benefiting from Google’s security and database solutions. Google Cloud CEO Thomas Kurian stated, “Anthropic and Google Cloud share a vision for developing AI boldly yet responsibly, ensuring secure and innovative AI solutions.”
This rapid evolution and scrutiny highlight the dynamic landscape in which AI startups operate, as they navigate growth, partnerships, and regulatory challenges.
Ranking
